one important note i like to make, a smartphone is a bad solution. people don’t want to mess around with it, they like it more when everything happens automaticaly. So from my point of view you have to build a platform that starts to learn the people and starts to predict there wishes. The google concept 
it works
Wow… You sure covered a lot in your long response.
There is a big difference between the theoretical/technical world and the practical one.
While your calculations and justifications regarding web-based (“Cloud”) services are technically correct, with full respect, I think you have seriously underestimated the real-world risks.
To be succinct, this is what I perceive as reality:
- The underlying cost of Cloud storage and compute power is irrelevant. What is relevant is that the Consumer can be easily trapped (locked) into paying whatever subscription price the Vendor demands; or be faced with abandoning at least some of their hardware sunk costs (the SmartHub) and all of their configuration effort (node joins, SmartApp configurations, Scenes, any SmartApps created using the proprietary API, ...).
<li>Likewise, the Consumer is at the mercy of the Vendor with respect to the quality of service (overall stability, security breaches [BTW: Dropbox is VERY unsecure], feature drops or major changes, bugs, outages). Far too many Vendors do NOT offer multiple channels (alpha, beta, stable) for their Clouds, and have far too frequent updates which can easily break key functionality. In essence, the Consumer is forever in the role of BETA TESTER. This is unacceptable for critical environments, and at the very least, frustrating for everyone else, particularly non-tech Consumers. In the context of SmartThings, failures can have far-reaching implications. The most obvious concern is a security breach. Over the years, Yahoo, Twitter, etc., etc., have had major security breaches, and the consequences difficult to measure (AP's Twitter account was hacked and that caused a stock market flash-crash -- $BILLONS LOST! A SmartThings security breach will (and it will happen) allow physical break-ins into homes as their SmartLocks are unlocked and their alarm systems are pre-disabled.</li>
<li>Finally, the Consumer has ZERO protection against dissolution of the Provider Enerprise (i.e., via bankruptcy or merger). Consider the uncountable number of services that Google has dropped with very little notice. Consider how nervous people are about Yahoo! purchasing Flickr, Tumbler, Astrid, ... The Astrid Task Management product is particularly relevant to me. I used this through over a year of unstable releases. The Android App alone has MILLIONS of installs, and 60,000 reviews. I purchased a subscription to premium services and this product became a critical part of my daily life. Then they sold out to Yahoo, and have given a strong (but very uncertain) indication that the product will be discontinued. I have hundreds of pieces of data stored in this tool that are not easily ported to a competitor. Regardless of the benefits of Astrid being "Cloud Based", I always was, and will always be, highly leary; and at the time I chose Astrid, I should have listened to my doubts and used non-SaaS application(s) instead (Outlook, etc.).</li>
In summary: The web-based/Cloud/SaaS is great in theory, but reality is that failures of this model are manifesting daily.
What’s the solution? Non-proprietary clouds are a good start… (OpenStack, Piston Cloud). But I fear there are too many embedded bad habits that the industry refuses to shake: Not surprisingly because these are quite profitable!
NB: SmartThings absolutely WILL sell out to the highest-bidder, if someone should care to buy them; even (and with non-trivial probability) if the sole purpose of buying them is to shut them down to destroy their competition, or to acquire their Customer Base and force these Consumers into upgrading their hardware or pay higher subscription costs.
You are right,
However one note i like to make: i want it all to be open source. In that way the costumer can control everything and there is no real owner. There will be no risk of selling and data loss. It are the consumers (or the small part of consumers that are willing, about 1%) that can alter applications if needed and make them stable. So instead of having some control at home the community will have full control.
The security aspect is true, however. I thought about this problem and I have a nice solution, 2 stage security. The ‘lock’ detects the user and recognizes him, this is the first stage of security, the second one is a pin code or rfid card that you need to use. Using only an encrypted RFID card is more secure than a lock. This security is hard-programmed and can be altered over the web unless the user gives his permission. It will be the community that can alter the way of working with it in consensus.
So, my point of view, you are right, businesses suck at this, open source is the way to go. This is the technology of the community.
What about when you lose the cloud? No access, no anything. Localized processing at least allows you to continue working on your projects, cloud or no.
it restricts you, integration of devices would be harder and you are limited in your calculation power. For simple functions it would suffice like turning on the light and things like that, but once you plan to let the house work on its own, it becomes harder. I plan on integrating the basic functions offline. The advanced functions online. Would you suggest working offline as much as possible?
I thought about it for a minute and came up with this solution: you make a basic platform, fully based on zigbee technology. Easy to develop, easy to deploy. With it comes and very basic zigbee remote and your central node has enough power to calculate the users position, estimate behavior and has a display with all the options. You can’t connect with it using your smartphone. To connect with it, you need to connect your device with the internet, your phone will have an app that will communicate with your house over the internet. The internet will analyze your data, compare it with others and uses the result to improve your home system. Apps and functions are stored local and run local, so you have all functionality offline.
Has been a very interesting read. I am not a programmer nor do i know about the hardware side of things - but am an enthusiast who can’t wait for home automation at lower prices. One thing I like about a majority of the software/rules happening locally is the speed component. Currently when i walk into the path of a sensored light, the light comes on very quickly. I don’t have my smartthings yet, but based on the videos I have seen, it looks as though there is a delay of about 2 seconds. For many connected scenarios this delay isn’t a problem, but for automatic lighting there will be many situations where this will be a problem. If due to my ignorance in this I am wrong, please let me know.
I also think it is likely that SmartThings may be (or wish to be) bought out, or otherwise may need to become truly open and possibly reassess their financial model (unless they plan on only funding via hardware). With the likes of Microsoft actively developing for the center of the home, a device as powerful as the Xbox One on the way, it looks as thought it could be a good bridge between the local and the cloud processing for Home Automation. Apps/scenes etc could be gathered in the cloud and downloaded to the local box. With the processing power on board there should be more than enough head room to have it all happening locally if chosen - with the option to remotely control the home as well. As far as security is concerned the sophistication of the new connect looks as though it could handle several biometric indicators as far as security is concerned. But again, I am not a programmer and am only guessing what is possible. Sorry if I have taken this very interesting topic off topic.
I’m glad that you’re so interested in this topic. The delay can be a problem, however I suspect that it can be limited to less than 0.5s, that is in most respects more than fast enough. Using the xbox ore something alike to power it all is impossible. The reason why is because Microsoft has a main goal playing, not home automation. The system uses too much power and has too few breakout options to ever become the center of the home automation platform. However, because you use the cloud, you can communicate with your platform using the internet. So indirect, you can control everything using any internet enabled device. In this way, the xbox can become you main control panel if you wish it to be so.
Welcome to the discussion, @Saxnix …
There is definitely some latency (lag delay in response) in SmartThings at the moment, but not sure where in the system this lives (remember there are multiple steps for every action: hub receives message from sensor or trigger, confirms receipt, sends to SmartCloud, Cloud processes SmartApp logic, sends commands to hub, hub sends command to light, light confirms…
I’m sure this will be analyzed to see where the delays are occurring. It is likely not the internet time causing the delay in most cases (the internet is surprisingly fast: Consider Google anticipating search results, and consider Skype which must relay your voice in real-time).
SmartThings does have a piece in their strategy for running some logic locally (they call it “wiring”, I think), and that will be helpful to solve a few problems with basic functions.
But I have listed many other concerns with a SmartThings “owned” Cloud.
I completely understand what you are both saying about the speed of the internet, but somewhere in the process (from what I have seen in a couple of videos), there is a lag which at the moment makes using it for automatic lighting a problem in some scenarios.
With the xbox one, I have read several articles that indicate that Microsoft may be looking to extend their automation interests in the future. They developed Microsoft Home, and are currently looking at updating that to a newer version (currently know as Home 2.0 - but that would not be its final name), which could find a platform on the Xbox One.
Here is one of the articles I read. Sorry the link function wasn’t working for me in Chrome.
I understand the cost is going to be around $600 for the One and I think the pricing on that is not too bad if you were able to use it for both a gaming machine as well as home automation - if it meant quicker responses/security advantages etc, etc.
I read the thread concerning a SmartThings owned cloud and have to agree with you and many of the posters there. At this point I am figuring that as long as the “Things” I buy have the ability to be used with other systems in the future, I can’t really go wrong. If the ST environment is it for a year or 2, and then some other better option/hub comes along and I go across to that, so be it. Sure there is time in learning a new system and setting things up, but if you are a hobbyist, it isn’t the end of the world. I certainly would hope that I could have everything running locally with a backup and access from the cloud. But do I want all my voice/video commands being uploaded to the cloud for the processing to happen - no. I would prefer it if whatever hub is used is doing all that locally and only going to cloud for upgrades, or possibly for very complex tasks/rules.
Regarding the actual topic of this thread, Marteen I think it would be awesome if there were some sort of in premises way of automating the home. As I wander around the home doing various things I would like to be able to enter and exit rooms and have the system sense/see/hear me and adjust lighting/music/temp etc as I like it, and also for the system to be able to do these same things for other members of the household. That is the dream and it will be an interesting journey getting there and who knows how long it will take, but either way I am looking forward to it and have most definitely been encouraged by the leaps and bounds that have been happening in hte recent year - SmartThings, UBI, UBE, LIFX, PlantLink - just to name a few.
You definitely see the possibilities: Hardware, even stuff “powerful enough” to run the full intelligence of most home automation tasks, really isn’t very expensive. I’m not sure exactly how powerful though – since it must multi-task (react to sensors throughout the home, and outside events, timers, etc.)… So it would be great if we could run our own “personal Clouds” with open-source software (non-proprietary APIs), just in case the SmartThings’s SmartCloud is slow, unstable, insecure, sold-off, …
I think, however, that SmartThings expects to minimize latency delay in the product. As such, if you have a reproducible case with a clear delay (1/2 second?) between event and action; I would suggest sending an email to support@SmartThings.com and they will probably go through the process of low-level debug tracing all the steps with you.
There are lots of possible reasons for lag time; but I think a lot of these will be improved as the system is tuned and scaled up.
It is good to hear that the latency should be less than a second - so mightn’t be a problem at all. I am only going by what I have seen in videos. As I am in Australia they don’t actually have SmartThings certified yet so I am left just following everyone else’s experiences, and continuing to research and plan 